Journal of the American Mosquito Control Association, 12(2):293-294, 1996 Copyright © 1996 by the American Mosquito Control Association, Inc. EVALUATION OF THE EFFICACY OF 3% CITRONELLA CANDLES AND 5% CITRONELLA INCENSE FOR PROTECTION AGAINST FIELD POPULATIONS OF AEDES MOSQUITOES L. ROBBIN LINDSAY, GORDON A. SURGEONER, JAMES D. HEAL and G. JAMES GALLIVAN Department of Environmental Biology, University of Guelph, Guelph, Ontario NIG 2WI Canada ABSTRACT. We assessed the efficacy of 3% citronella candles and 5% citronella incense in protecting subjects from bites of Aedes spp. under field conditions. The study was conducted in a deciduous woodlot in Guelph, Ontario, Canada from July 26 to August 10, 1995. Eight subjects, dressed identically, were assigned to one of 8 positions on a grid within the study area. Two citronella candles, 2 citronella incense, 2 plain unscented candles, or no candles (i.e., nontreated controls) were assigned to 2 positions on the grid each evening. Subjects conducted 5-min biting counts at each position and performed 16 biting counts per evening. On average, subjects received 6.2 ± 0.4, 8.2 ± 0.5, 8.2 ± 0.4, and 10.8 ± 0.5 bites/ 5 min at positions with citronella candles, citronella incense, plain candles, and no candles, respectively. Although significantly fewer bites were received by subjects at positions with citronella candles and incense than at nontreated locations, the overall reduction in bites provided by the citronella candles and incense was only 42.3 and 24.2%, respectively. Candles containing oil of citronella are sold commercially in the United States and Canada, with some manufacturers claiming that these products "reduce the annoyance of biting in- sects," specifically mosquitoes. However, with the exception of unpublished data cited in Curtis (1986) that suggested that citronella candles do not appreciably reduce the biting rates of mos- quitoes, quantitative studies of the repellency provided by citronella-based candles or incense are lacking. The purpose of this study was to assess the efficacy of commercially available 3% citronella candles and 5% citronella incense to protect people from bites of Aedes spp. mosqui- toes under field conditions. The study was conducted in a deciduous woodlot in Guelph, Ontario (43°45'N, 80°20'W). This woodlot, which is surrounded by large snowmelt pools where large numbers of imma- ture Aedes spp. develop has been described pre- viously (Surgeoner and Heal 1992, Heal et al. 1995). Eight subjects (4 ? $; 4 3 6) were used in this study, which was conducted on 8 nights (July 28, 31 and August 1, 2, 7, 8, 9, 10, 1995). An additional person timed biting counts and re- corded weather conditions. The subjects wore headnets, white cotton gloves, and green cov- eralls with the sleeves rolled up to the elbows. They were not allowed to wear repellent. Biting counts were initiated by 1900 h each night. The subjects were assigned to one of 8 positions on a grid within the study area at the beginning of the evening and rotated through all 8 positions twice each night. Positions were sep- arated by at least 10 m. Each treatment was placed at 2 positions on the grid each evening. The treatments were citronella candles, citronella incense, plain unscented candles, and no candles (i.e., nontreated controls). Two citronella candles, 2 plain candles, or 2 citronella incense were placed at each treated position on top of 35-cm plastic stands 1 m apart. A plastic lawn chair was placed between the plastic stands and subjects conducted biting counts while seated on the lawn chairs. Treatments were assigned to positions on the grid such that each treatment was at each po- sition twice during the 8-night evaluation. Biting counts were made during 5-min peri- ods at each position. The subjects aspirated all of the mosquitoes biting both exposed forearms into 150-ml clear plastic vials. The subjects re- corded the number of mosquitoes captured and then moved to the next position on the grid. Thus, subjects made 4 biting counts at each treatment, each night. Mosquitoes collected were pooled each night and 30/night were ran- domly selected and identified using keys of Wood et al. (1979). Ambient temperature, rela- tive humidity, wind speed, and wind direction within the study site were recorded at =30 min intervals during the biting counts. The data were analyzed using a 4-factor anal- ysis of variance with subject, treatment, position on the grid, and night as the independent vari- ables. Because the number of mosquitoes col- lected per 5-min biting count varied widely be- tween subjects, and the variance was correlated with the mean, these data were log 10 transformed prior to the analysis. Subject, position on the grid, and night were random variables and treat- ment was a fixed variable. The analyses were completed using Statistical Analysis Systems 293 294 Journal of the American Mosquito Control Association Vol. 12, No. 2 Table 1. Mean number ± SE of mosquitoes collected per 5-min biting count (n = 128) at positions with 3% citronella candles, 5% citronella incense, plain candles, or nontreated controls. Treatment Parameter Citronella candles Citronella incense Plain candle Mean ± SE 1 Percent reduction 2 6.2 ± 0.4A 42.3 1.2 ± 0.5B 24.2 8.2 ± 0.4B 23.1 No treatment 10.8 ± 0.5C 1 Means followed by the same letter are not different at P < 0.05; Student-Newman-Keuls test. 2 Percent reduction calculated as: [(no. biting subjects at nontreated positions - no. biting at treated positionsWno. biting at nontreated positions)] X 100. version 6.04 (SAS Institute Inc., Cary, NC). Per- cent reduction in the number of bites provided by each treatment was also calculated as: [(no. biting subjects at nontreated positions - no. bit- ing at treated positions)/(no. biting at nontreated positions)] X 100. Each night 800-1,300 mosquitoes were col- lected. Ten mosquito species were identified from 240 individuals subsampled throughout the study. The species and percent composition were: Aedes euedes Howard, Dyar and Knab (59.2%), Ae. vexans (Meigen) (12.5%), Ae. fitchii (Felt and Young) (11.2%), Ae. trivittatus (Coquillett) (10.4%), Ae. excrucians (Walker) (2.5%), Ae. canadensis (Theobald) (2.1%), Ae. stimulans (Walker) (0.8%), Ae. implicates Vockeroth (0.4%), Coquillettidia perturbans (Walker) (0.4%), and Anopheles quadrimaculatus Say (0.4%). Ambient air temperatures ranged from 19 to 26°C during biting count evaluations. Wind was less than 5 km/h during all nights except for <30 min on August 1 when gusts up to 8 km/h were recorded. Regardless of the wind direction, the subjects were always exposed to the smoke of at least one candle or incense. Light rain fell for < 10 min on August 1; however, it did not rain on any other evening. None of the meteorological param- eters measured would have deterred host-seeking activity by mosquitoes and therefore did not affect the results obtained. Subjects received significantly fewer bites (P < 0.001) at positions with citronella candles compared with the other 3 treatments (Table 1 ). Significantly more mosquitoes (P < 0.001) were collected at nontreated positions than ones with citronella incense and plain candles and biting counts did not differ significantly (P > 0.5) be- tween these 2 treatments (Table 1). Although significantly fewer mosquitoes were collected at treated positions compared with nontreated ones, the overall percent reduction provided by the cit- ronella candles, citronella incense, and plain candles was only 42.3, 24.2, and 23.1%, respec- tively. Surprisingly, plain candles reduced the biting activity of mosquitoes, presumably be- cause the light, heat, carbon dioxide, and mois- ture produced drew mosquitoes away from the subjects. Considering the reduction in mosquito biting activity produced by the act of burning candles, the addition of 3% citronella to candles further decreased biting activity by only 19.2%. As in previous studies with woodsmoke (Snow et al. 1987), citronella candles or incense were ineffective for reducing the biting pressure of mosquitoes, and their use by the general public should be discouraged. Although one manufac- turer (S.C. Johnson & Son, Limited, Brantford, Ontario) specified that the candles should be placed =45 cm apart, for practical and safety rea- sons we spaced the candles and incense 1 m apart. In general the directions for use of these products are vague or nonexistent and consumers are not explicitly told how many candles must be used to produce adequate levels of protection from mosquitoes. Increasing the number of can- dles or incense per unit area may have increased the efficacy of these products; however, it is un- likely that consumers would tolerate the exces- sive amounts of smoke and odor that would have to be generated to produce this effect. We thank the Ontario Ministry of Agriculture and Food and Rural Affairs for continued in- frastructural support. We also thank the subjects who participated in this study. REFERENCES CITED Curtis, C. F. 1986. Fact and fiction in mosquito at- traction and repulsion. Parasitol. Today 2:316-318. Heal, J. D„ G. A. Surgeoner and L. R. Lindsay. 1995. Permethrin as a tent treatment for protection against field populations of Aedes mosquitoes. J. Am. Mosq. Control Assoc. 11:99-102. Snow, R. W., A. K. Bradley, R. Hayes, P Byass and B. M. Greenwood. 1987. Does woodsmoke protect against malaria? Ann. Trop. Med. Parasitol. 81:449-451. Surgeoner, G. A. and J. D. Heal. 1992. Evaluation of three deet (N, A^-diethyl-m-toluamide) formulations, Avon-Skin-So-Soft and the citrosa plant as repellents against spring Aedes spp. pp. 151-153. In: Agric. Can- ada Publ. Pest Manag. Res. Rpt. Ottawa, Canada. Wood, D. M., P. T. Dang and R. A. Ellis. 1979. The insects and arachnids of Canada. Part 6: The mos- quitoes of Canada. Agric. Canada Publ. 1686.